Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
10
,
No.
2
,
A
pr
il
2020
, p
p. 13
87
~
1397
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v10
i
2
.
pp1387
-
13
97
1387
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om/i
nd
ex
.ph
p/IJ
ECE
A hybrid
approach fo
r schedul
ing
appli
cations
in cl
oud
co
mp
uti
ng envir
onment
Ah
med
Subhi
Ab
d
alk
afor
1
,
Khattab
M.
A
li
A
lhee
ti
2
1
Care
er
Deve
lop
m
ent
Center, Unive
rsit
y
o
f
Anba
r,
Anbar
,
Ir
aq
2
Com
pute
r
Netw
orking
S
y
st
ems
Depa
rtment, Co
m
pute
r
Scie
n
ce
s
and
In
form
at
ion
Technol
og
y
Col
le
ge
,
Univer
sit
y
o
f
An
bar
,
Anb
ar,
Ira
q
,
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
J
un
28
, 2
019
Re
vised
O
ct
9
,
2019
Accepte
d
O
ct
15
,
2019
Cloud
computin
g
play
s
an
important
rol
e
in
our
dai
l
y
l
ife.
It
ha
s
dire
ct
an
d
positi
ve
impact
on
share
and
updat
e
data,
knowl
edge
,
storag
e
an
d
scie
nti
fi
c
resourc
es
bet
w
e
en
var
ious
reg
i
ons.
Cloud
computing
per
form
a
nce
he
avil
y
base
d
on
job
sc
hedul
ing
a
lgor
ithm
s
tha
t
are
uti
li
z
ed
for
queue
wait
ing
in
m
oder
n
scie
nti
f
ic
applic
at
ions.
The
rese
arc
h
ers
are
consid
ere
d
c
lou
d
computing
a
p
opula
r
pl
at
form
for
new
enf
or
ce
m
ent
s.
The
se
sche
duli
n
g
al
gorit
hm
s
hel
p
in
design
eff
icie
nt
queue
li
sts
in
cl
oud
as
well
as
th
e
y
p
l
a
y
vit
al
ro
le
in
re
duci
ng
wai
ti
ng
for
proc
essing
ti
m
e
in
cl
oud
computing.
A
novel
job
sch
edul
ing
is
propo
sed
in
th
is
pape
r
to
enh
ance
per
f
orm
anc
e
of
cl
oud
computin
g
and
red
uce
del
a
y
ti
m
e
in
queue
wait
in
g
for
jobs.
The
proposed
a
lgori
thm
tries
t
o
avoi
d
som
e
signifi
c
ant
ch
all
enge
s
that
throt
tle
from
de
vel
oping
appl
i
cations
of
cl
oud
c
om
puti
ng.
How
eve
r,
a
sm
art
sche
duli
ng
t
ec
h
nique
is
proposed
in
our
pap
e
r
to
improve
per
form
ance
proc
essing
in
c
lo
ud
appl
icat
ions.
Our
expe
rimental
result
of
th
e
pr
oposed
job
sche
duli
ng
al
gor
it
hm
show
s
tha
t
the
proposed
sc
hemes
poss
ess
o
utsta
ndin
g
enha
nc
ing
r
at
es
with
a
red
u
ct
ion
in
wai
ti
ng
ti
m
e
f
or
jobs i
n
queu
e li
st.
Ke
yw
or
d
s
:
Cl
oud
c
om
pu
ti
ng
Task sc
he
duli
ng alg
ori
thm
p
r
i
o
r
i
t
y
Copyright
©
202
0
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Ah
m
ed
S
ubhi
Abdalka
f
or
,
Ca
reer De
velo
pm
ent Center,
Un
i
ver
sit
y o
f Anba
r,
Anba
r,
Ir
a
q
.
Em
a
il
:
ah
m
ed.
abd
al
kafor@
uoan
ba
r.
e
du.iq
1.
INTROD
U
CTION
Nowa
days,
m
od
er
n
te
ch
no
l
ogy
is
witnessing
a
wide
ra
ng
e
of
dev
el
op
m
ent
in
sci
entifi
c
app
li
cat
io
ns
and
resea
rc
h
a
rea
s
.
Cl
oud
c
om
pu
ti
ng
is
c
onside
red
one
of
t
he
m
os
t
i
m
p
or
ta
nt
resea
rc
h
fiel
d
s
t
hat
rec
ei
ved
a
lot
of
at
te
ntion
f
r
om
dev
el
op
e
rs
an
d
desi
gn
e
rs.
It
is
an
inn
ovat
ive
te
chnolo
gy
that
us
es
cent
r
al
rem
ote
serv
e
rs
a
nd
i
ntern
et
netw
orks
to
s
ha
re
in
f
orm
at
ion
/a
pp
li
ca
ti
on
s
a
nd
st
or
e
im
po
rtant
data
to
bec
om
e
avail
able
for
us
ers
[
1,
2].
Cl
oud
c
om
pu
ti
ng
al
lows
co
nsum
ers
t
o
s
har
e
a
nd
a
ccess
f
or
dail
y
app
li
cat
ions/
sensiti
ve
inf
or
m
at
ion
wi
thout
c
oord
i
na
ti
on
a
nd
syn
ch
ronizat
ion
m
od
e.
The
se
a
pp
l
ic
at
ion
s
pro
vide
im
po
rtant
da
ta
an
d
con
t
ro
l i
nfor
m
at
ion
to
users
a
t suita
ble
ti
m
e
and w
it
ho
ut any
d
el
ay
[3,
4].
Cl
oud
c
om
pu
ti
ng
util
iz
es
co
m
pu
ti
ng
res
ou
rces
a
nd
net
w
ork
te
rm
inals
to
pr
ov
i
de
im
portant
data,
inf
or
m
at
ion
,
and
knowle
dge
for
us
e
rs
at
the
ap
pro
pr
ia
te
tim
e.
In
oth
e
r
words,
use
rs
can
acce
ss
the
sa
m
e
resou
rce
at
dif
fer
e
nt
tim
e
s
and
place
i.e.
s
ha
re
data.
T
he
m
ai
n
aspect
of
cl
oud
com
pu
ti
ng
does
not
nee
d
pre
-
est
ablished
c
om
m
un
ic
at
ion
f
or
any
request
from
resour
c
es
.
T
his
as
pec
t
m
akes
cl
oud
com
pu
ti
ng
oc
cup
y
a
la
rg
e
a
rea
of
sci
entifi
c
rese
arch
an
d
m
od
e
rn
a
ppli
cat
ion
s
in
our
daily
li
fe.
T
he
cl
ou
d
com
pu
ti
ng
pe
rm
i
ts
acce
ss
to
se
ve
ral
onli
ne
ser
vi
ces,
so
ci
al
dat
a,
inf
or
m
at
ion
and
res
ources
that
can
be
ut
il
iz
ed
fo
r
e
xc
han
ge
betwee
n
com
pu
te
r
de
vices
on
dem
and
.
I
n
order
to
al
loc
at
e
co
m
pu
ti
ng
resour
ces
qui
ckly
and
eff
ic
ie
ntly
,
resou
rce
al
loc
at
ion
ta
s
k
s
m
us
t
be
sc
he
du
l
ed
[5
-
8].
T
he
basic
str
ucture
of
cl
oud
c
om
pu
ti
ng
is
s
how
n
i
n
Fig
ure
1.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
2
,
A
pr
i
l 202
0
:
1387
-
1397
1388
Figure
1. Ba
sic
struct
ur
e
of cl
oud
c
om
pu
ti
ng
[
9]
Accor
ding
to
Figure
1,
data
stora
ge,
se
rv
e
r
s,
operati
ve
sy
stem
,
and
vi
rtu
al
iz
at
ion
interf
ace
are
m
a
in
com
po
ne
nts
f
or
any
cl
oud
c
om
pu
ti
ng
syst
e
m
[9
]
.
T
he
di
ff
ic
ulty
of
jo
b
sc
hedulin
g
r
equ
e
st
s
is
gra
du
al
ly
increase
d
in
cl
ou
d
syst
em
.
In
m
or
e
detai
l,
the
cu
rr
e
nt
al
gorithm
s
becom
e
par
al
yz
ed
in
f
ront
of
the
pool
request
of
cl
oud
web
a
pp
li
c
at
ion
s.
As
we
know,
sche
duli
ng
in
a
cl
oud
c
om
pu
ti
ng
env
ir
onm
ent
is
m
os
t
i
m
po
rtant
iss
ue
in
m
od
er
n
te
chnolo
gy.
T
he
ta
sk
sc
heduling
c
onsist
s
of
the
f
ollow
i
ng
:
workflo
w,
sta
ti
c
an
d
dynam
ic
cl
ou
d
ser
vice
sc
he
du
li
ng,
real
-
t
i
m
e
sched
ulin
g,
heurist
ic
and
oppo
rtu
nisti
c
load
bala
nci
ng
sche
du
li
ng
[10
,
11]
.
T
he
c
ompu
ti
ng
e
nvir
onm
ent
consi
sts
of
a
num
ber
of
sources
.
T
here
fore,
ta
s
k
sc
he
du
li
ng
is
necessa
ry
t
echn
i
qu
e
to
c
hoos
e
t
he
bes
t
resou
rces
s
ui
ta
ble
for
the
i
m
ple
m
entat
i
on
of
var
i
ous
ta
sk
s
.
The
e
ff
ic
ie
nt
s
cheduli
ng
sc
he
m
e
can
re
du
ce
respo
ns
e
ti
m
e,
low
c
onsu
m
ption
pow
er,
an
d
pro
vid
e
a
vaila
bili
ty
for
busy
res
ources.
I
n
ad
diti
on,
sche
duli
ng
m
et
ho
ds
ha
ve
the
abili
ty
to
al
locat
ing
com
pu
te
r
m
achine
s
with
the
le
ast
tim
e
t
o
accom
plish
dep
e
ndin
g
on
t
heir
pr
io
riti
es,
so
sc
hedulin
g
t
asks
a
re
esse
ntial
and
in
flue
nt
ia
l
i
n
this en
vir
on
m
ent. Fo
r
this
rea
so
n,
new sc
heduling al
gorith
m
s ar
e requested to
ove
rco
m
e this
prob
l
em
.
In
this
pap
e
r,
a
novel
sche
du
li
ng
al
gorit
hm
is
pr
op
os
e
d
to
im
pr
ov
e
the
perf
or
m
ance
of
cl
ou
d
com
pu
ti
ng
.
M
or
e
over,
it
has
the
a
bili
ty
to
re
du
ce
the
a
m
ou
nt
of
dela
y
in
qu
e
ue
w
ai
ti
ng
.
T
he
pr
opos
e
d
al
gorithm
in
this
resear
ch
pl
ay
s
i
m
po
rtant
ro
le
in
al
loc
at
ing
bala
nce
load
be
twee
n
diff
e
re
nt
resou
rces.
Ther
e
f
or
e,
the
resou
rce
-
sc
he
duli
ng in
cl
oud
com
pu
ti
ng
is
di
ff
ic
ult t
a
sks [
10,
11]
.
2.
RELATE
D
W
ORK
Re
centl
y,
seve
ral
sche
du
li
ng
al
gorithm
s
hav
e
been
pro
po
se
d
to
distri
bu
te
cl
oud
ap
plica
ti
on
s
.
All
of
the
se
al
gorith
m
s
ta
rg
et
to
re
du
ce
wait
ing
t
i
m
e
of
j
obs
in
qu
e
ue
li
st,
in
crease
ef
fici
en
cy
,
decli
ne
nu
m
ber
of
rep
eat
re
quest
and
balanc
e
be
tween
a
vaila
bl
e
resour
ces
.
I
n
this
pap
e
r,
we
pr
opos
e
d
ne
w
hy
br
i
d
sch
edu
li
ng
al
gorithm
to
e
nh
a
nce
perfor
m
ance
of
cl
oud
com
pu
ti
ng.
Sangwa
n
et
al
.
[12]
pro
posed
te
chn
iq
ue
t
o
im
pr
ov
e
rou
nd
-
r
obin
sc
hedulin
g
al
gori
thm
in
cl
oud
c
om
pu
ti
ng
en
vi
ronm
ents
by
re
du
ci
ng
the
av
e
rag
e
of
wait
ing
tim
e
and
total
arou
nd
ti
m
e
by
chan
gi
ng
the
ti
m
e
qu
a
ntu
m
by
keep
i
ng
al
l
the
pr
o
ce
sses
w
hen
ar
rivi
ng
i
n
CPU.
It
de
fines
a
verage
of
m
ean
(T
Q
po
werfu
ll
y)
after
sel
ect
in
g
the
m
ean
of
burst
tim
e
according
t
o
the
num
ber
of
processes
.
T
hi
s
pro
pose
d
al
gorithm
sel
ect
s
the
fi
rst
pr
oce
ss
the
n
CP
U
i
s
al
locat
ed
f
or
a
te
m
po
rar
y
per
i
od
accor
ding
to
the
m
ean
t
i
m
e
qu
a
nt
um
.
There
are
m
any
st
ud
ie
s
that
ha
ve
been
w
orke
d
on
ta
sk
sche
duli
ng.
Howe
ver,
the
r
e
is
sti
ll
a
cha
ll
eng
e
for
sch
edu
li
ng
al
gorit
hm
s,
su
ch
as
m
ini
m
u
m
respon
s
e
ti
m
e,
m
a
xim
u
m
thr
oughput,
be
st
po
ssi
ble
res
ource
util
iz
at
io
n,
a
nd
re
du
ci
ng
the
ov
e
rloa
d
[13].
Mitta
l
a
nd
Sin
gh
prese
nt
[
14
]
a
propose
d
sys
tem
to
i
m
pr
ove
the
al
gorithm
for
ta
sk
sc
he
duli
ng
i
n
cl
oud
com
pu
ti
ng
.
Th
e
auth
or
s
ca
n
m
od
ify
rou
nd
-
r
obin
by
add
i
ng
a
new
strat
egy
of
th
e
per
io
d
f
or
re
qu
e
st
tim
e
qu
antum
.
In
m
or
e
detai
l,
the
pro
po
s
e
d
al
gorithm
ob
ta
ins
m
ini
m
iz
es
wait
ing
ti
m
e
a
nd
m
axi
m
iz
e
CPU
usa
ge
.
K
um
ar
et
al
.
[15]
hav
e
t
ried
to
fin
d
the
best
slot
of
ti
m
e
qu
ant
um
in
ro
un
d
-
r
ob
i
n
al
gorith
m
to
avo
id
t
he
fr
e
qu
e
nt
co
ntext
s
witc
h.
In
te
ger
pro
gr
am
m
ing
is
util
i
z
ed
in
create
new
CP
U
local
sche
duli
ng
nam
ed
changea
ble
tim
e
qu
a
ntu
m
(CTQ)
t
o
ov
e
rc
om
e
too
la
rg
e/
to
o
li
tt
le
of
t
he
ti
m
e
quantum
.
More
over
,
it
has
the
abili
ty
to
giv
es
the
best
th
rou
ghput
rate o
f
the
syst
e
m
.
In
[16],
the
a
uthors
pro
pose
d
ne
w
sc
hem
e
for
re
duci
ng
so
m
e
of
the
draw
bac
ks
of
r
ound
-
robin
al
gorithm
as
well
as
getti
ng
bette
r
per
f
or
m
ance
of
CPU
at
the
sam
e
tim
e.
In
ot
her
words,
the
m
od
ific
at
ion
of
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
y
br
id
ap
pr
oach
for
sch
e
du
li
ng
applicati
on
.
..
(
A
hm
e
d S.
Ab
da
lk
afo
r)
1389
al
gorithm
plays
i
m
po
rtant
r
ol
e
in
reducin
g
aver
a
ge
of
wai
ti
ng
tim
e,
decli
ne
the
am
ou
nt
of
tu
rn
a
r
ound
ti
m
e
and
dec
reases
the
dif
ficult
of
sta
r
vatio
n.
Nayak
e
t
al
.
[17]
ha
ve
im
pro
ved
t
he
r
ound
-
r
ob
i
n
al
gorithm
by
orga
nizin
g
the
pr
ocesses
i
n
the
rea
dy
que
ue.
T
his
m
od
if
ic
at
ion
can
ove
rco
m
e
the
dr
a
wb
ac
ks
of
al
go
rith
m
by
set
opti
m
a
l
tim
e
qu
antum
and
sel
ect
the
sh
ort
est
bu
rst
tim
e.
In
[
18]
,
a
sta
rv
at
io
n
op
ti
m
iz
ing
a
lgorit
hm
is
i
m
pr
oved
to
fix
one
of
the
c
omm
on
pr
oble
m
s
that
occu
r
in
shortest
job
first
al
gorith
m
.
This
m
od
if
ic
at
ion
play
s
vital
ro
le
in
enh
a
ncin
g
CPU
and
in
put/
ou
t
pu
t
res
ources
re
quirem
ents
that
hav
e
a
direct
and
posit
ive
ref
le
ct
io
n
on
overall
perf
or
m
ance
.
It
can
in
crease
thr
ough
pu
t
rate,
re
duc
e
aver
a
ge
wait
ing
tim
e
and
de
cl
ine
tur
naroun
d
ti
m
e
in
cl
oud
e
nv
i
ronm
ent.
Wh
e
n
res
ourc
es
are
overl
oa
ded
the
al
gori
thm
has
re
dir
ect
ed
al
l
resou
rces
in
vi
rtual
m
achines.
The
virtu
al
m
achine
in
t
his
syst
e
m
aims
to
get
be
tt
er
res
ponse
an
d
proc
e
ssin
g
tim
e.
Gh
an
bar
i
et
al
.
[
5],
have
pro
posed
a
ne
w
al
gorithm
base
d
on
t
he
a
naly
ti
cal
hierarc
hy
proce
ss
na
m
ed
PJSC.
It
f
ocus
es
on
th
ree
m
at
te
rs
wh
ic
h
a
re
co
ns
ist
ency,
com
plexity
,
a
nd
m
akes
-
pa
n.
Con
sist
ency
m
eans
adjustin
g
fun
dam
enta
ls
of
c
om
par
ison
m
at
ri
x
based
on
sch
edu
li
ng
of
pr
io
rity
al
go
rithm
.
Wh
e
n
c
om
plexity
is
the
num
ber
of
j
obs
as
well
as
it
is
cal
culating
the
pri
ori
ty
vector
s
of
e
valu
at
ion
m
at
rixes.
Ag
a
rw
al
et
al
.
[19]
,
hav
e
prese
nted
a
ge
ner
al
pri
or
it
y
al
gorith
m
fo
r
ef
fecti
v
e
i
m
ple
m
entation
f
or
res
our
ce
al
locat
ion
in
cl
ou
d
com
pu
ti
ng
. I
t
pro
vid
es
a
bette
r
perform
ance
after
app
ly
in
g
in
cl
oud
sim
to
olk
it
env
i
ronm
ents
an
d
com
par
is
on
with
rou
nd
-
r
ob
in and
first c
om
e first serves
al
gorithm
s.
A
novel
hybri
d
sc
he
du
li
ng
a
lgorit
hm
is
p
r
opos
e
d
i
n
this
pap
e
r
t
o
ov
e
rc
om
e
co
m
m
on
pro
blem
s
of
the
curre
nt
scheduli
ng
al
gor
it
h
m
s.
It
play
s
a
direct
an
d
po
sit
ive
im
pact
on
cu
rr
e
nt
m
et
ho
ds
via
r
edu
ci
ng
wait
ing
ti
m
e,
t
urnaro
und
tim
e
and
inc
reasi
ng
perform
ance
of
cl
oud
c
om
pu
ti
ng
syst
em
.
This
al
go
ri
thm
has
the
abili
ty
to
e
sta
blish
bala
nc
e
for
avail
able
resour
ce
s.
T
hus,
it
will
be
pro
vid
i
ng
s
uffic
ie
nt
trans
par
e
nc
y
fo
r
cl
oud
us
e
rs
in
diff
e
ren
t
reg
i
on
s
.
T
he
obj
e
ct
ive
of
this
pa
per
is
to
pro
po
s
e
a
ne
w
al
gorithm
to
enh
anc
e
perform
ance
of
the
r
ound
-
robin
al
go
rithm
via
dec
reasi
ng
am
ou
nt
of
w
ai
ti
ng
ti
m
e
and
tu
r
naroun
d
a
ver
a
ge
tim
e.
This
ench
antm
ent
is
ac
hieve
d
by
ad
di
ng
pr
i
or
it
y
fiel
d
to
arr
i
val
tim
e
of
jo
b
proces
ses
wh
et
her
e
qual
or
diff
e
re
nt.
3.
CLOUD
COMP
UTING
The
jo
b
sche
duli
ng
al
gorith
m
s
in
cl
ou
d
c
om
pu
ti
ng
a
re
cl
assifi
ed
into
two
cat
eg
or
ie
s
:
batch
m
od
e
heurist
ic
al
go
r
it
h
m
(BMHA)
and
onli
ne
m
od
e
heurist
ic
al
gorithm
(O
MHA
)
[
5].
The
BM
HA
colle
ct
s
and
qu
e
ue
d
in
a
set
fo
r
a
rr
i
ve
jo
b
processes
at
th
e
wait
ing
li
st.
The
tim
e
is
fixed
in
this
al
gor
it
h
m
fo
r
jo
bs
in
cl
o
ud
com
pu
ti
ng
.
Fi
r
st
co
m
e
first
s
erv
e
d
al
gorith
m
is
con
side
re
d
one
of
the
a
pp
li
cat
io
n
s
of
BM
HA
.
I
n
ad
diti
on
,
rou
nd
-
r
obin,
Ma
x
-
Min
,
an
d
Mi
n
-
Mi
n
al
gorithm
s
are
al
so
co
ns
i
der
e
d
app
li
cat
ions
of
BM
H
A.
W
her
eas
,
OMH
A
fetc
h
es
an
d
processe
s
jobs
f
ro
m
queue
wait
ing
d
ir
ect
ly
[8
,
20,
21]
.
T
he
BM
H
A
is
c
on
si
der
e
d
on
e
of
the
m
os
t
su
it
able
al
gorithm
s
for
t
he
cl
ou
d
com
pu
ti
ng
e
nv
iro
nm
ent.
In
t
his
a
ppro
ac
h,
al
l
arr
ive
d
j
ob
s
ap
ply
on
e
of t
he
m
eth
ods
bel
ow
[
22]
:
3.1.
Fir
st
c
om
e
first
serve
d
(
FCFS)
1
It
i
s
al
so
nam
ed
First
i
n
First
Ser
ves
(
FI
F
S)
wh
e
n
the
job
proces
s
that
ar
rived
at
qu
e
ue
wait
ing
first
will
be
se
rv
e
d.
T
he
se
rv
es
perform
by
order
in
read
y
qu
e
ue,
an
d
t
he
y
do
not
pro
cess
sho
rt
or
la
rg
e
.
The
process
i
n
the
que
ue
m
us
t
be
wait
in
g
unti
l
al
l
pr
oces
s
es
are
c
om
plete.
T
he
proces
s
es
wait
ing
m
ore
ti
m
e
to term
inate
f
r
om
r
eady q
ue
ue
, so the
w
ai
ti
ng a
nd tu
rn
a
rou
nd are
quit
e h
i
gh [2
3].
3.2.
Min
-
mi
n
algorithm
T
his
al
gorith
m
determ
ined
al
l
ta
s
ks
an
d
a
rr
a
ng
e
d
it
ab
out
la
rg
e
a
nd
s
m
al
l,
the
s
m
al
l
ta
sk
s
will
be
execu
te
d,
fir
stl
y
w
hile
the
lar
ge
ta
sk
delay
s
fo
r
a
lo
ng
ti
m
e.
The
dr
a
w
back
of
this
al
gorithm
is
to
assig
n
the sm
al
le
r
pr
oc
esses to
the
re
so
urces
w
it
h c
om
par
at
ively
hi
gh
e
r
c
om
pu
ta
ti
on
al
powe
r
[
24]
.
3.3.
M
ax
-
mi
n
algorithm
T
his
al
gorith
m
al
so
dep
e
nds
on
sm
al
l
a
nd
la
r
ge
ta
s
ks.
H
ow
e
ver,
it
changes
to
t
he
Min
-
Min
al
gorithm
wh
e
nev
e
r
t
he
ta
sk
was
to
o
la
rg
e
.
It
will
be
exec
uted
a
s
s
oon
a
s
an
d
delay
a
sm
a
ll
ta
sk
for
long
-
tim
e [
25
]
.
3.4.
R
ou
n
d
-
r
obi
n
a
l
go
ri
th
m
I
t
is
s
uitable
f
or
cl
ou
d
c
om
pu
ti
ng
ap
plica
ti
ons.
I
n
ad
diti
on,
t
he
m
os
t
effe
ct
ive
ti
m
e
-
sh
arin
g
syst
em
s
,
i
t’s si
m
plest an
d
m
os
t widely
us
e
d
pro
portio
nal sh
are sc
he
duli
ng
alg
ori
th
m
[
11
]
. Th
e a
ve
rag
e
wait
time
(
W
T
)
and
a
ver
a
ge
tu
rn
e
d
ar
ound
ti
m
e
(TA
T)
a
re
head
po
i
nts
on
syst
em
,
so
ro
un
d
-
robin
de
sign
e
d
to
giv
e
so
m
e
bette
r
res
ponsi
ve
but
the
w
or
st
a
ver
a
ge
outp
ut
pa
ram
eter
s
(
WT
,
TA
T),
i
t'
s
no
t
ef
fici
ent
en
ough
to
be
e
m
plo
ye
d
i
n
re
al
-
tim
e syst
e
ms [5].
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
2
,
A
pr
i
l 202
0
:
1387
-
1397
1390
3.5.
Pri
orit
y
s
cheduli
ng al
gori
th
m
I
n
this
al
gorithm
,
the
FCFS
proces
ses
sc
he
du
l
e
d
is
a
ppli
ed
wh
e
n
j
obs
hav
e
eq
ual
-
pri
or
it
y.
It
ha
s
a
ve
ry im
po
rtant ef
fect o
n pe
r
form
ance o
f
cloud c
om
pu
ti
ng in
c
urren
t a
ppl
ic
at
ion
s.
3.6.
Th
e
Sh
or
t
est
-
Job
-
Fir
st
(
SJ
F
) alg
orith
m
I
t
is
a
s
pecial
case
of
the
ge
ner
al
pri
o
rity
sche
duli
ng
al
gorithm
[6
]
,
i
n
c
om
pu
te
r
sc
ie
nce,
queu
e
pr
i
or
it
y
as
the
kind
of
abstra
c
t
data,
wh
ic
h
is
li
ke
a
qu
eue
r
egu
la
rly
or
str
uctu
re
of
the
da
ta
sta
cked
but
w
h
e
re
add
it
io
nally
ea
ch
el
e
m
ent
has
a
"pr
iority
"
associat
ed
wit
h
it
.
This
al
gorithm
'
s
wo
rk
wh
e
n
each
pro
cess
has
high
pri
ori
ty
i
s
execu
te
d
be
f
or
e
a
lo
w
pr
i
ori
ty
pr
oce
ss.
I
f
two
processes
hav
e
t
he
sam
e
pr
io
rity
,
the
y
are
execu
te
d
acc
ordin
g
to
t
he
ir
order
i
n
the
queu
e
[26
]
.
T
he
co
m
m
on
ty
pes
of
s
che
du
li
ng
al
gorithm
s
that
util
iz
ed
in clo
ud
com
puti
ng is s
how
n i
n
Fig
ure
2
.
Fig
ur
e
2. Ca
te
gories
of
s
che
duli
ng
a
lg
or
it
hm
W
o
r
k
f
l
o
w
s
t
e
c
h
n
i
q
u
e
i
s
u
t
i
l
i
z
e
d
i
n
t
h
i
s
p
a
p
e
r
t
o
d
e
t
e
r
m
i
n
e
t
h
e
o
p
t
i
m
a
l
d
i
s
t
r
i
b
u
t
i
o
n
o
f
t
h
e
i
r
j
o
b
p
r
o
c
e
s
s
e
s
t
o
a
v
a
i
l
a
b
l
e
r
e
s
o
u
r
c
e
s
w
h
e
t
h
e
r
c
o
m
p
u
t
a
t
i
o
n
a
l
n
e
t
w
o
r
k
.
A
p
p
l
i
c
a
t
i
o
n
s
t
r
u
c
t
u
r
e
s
i
n
w
o
r
k
f
l
o
w
s
s
t
a
g
e
a
r
e
a
n
a
l
y
z
e
d
t
o
s
e
l
e
c
t
a
p
p
r
o
p
r
i
a
t
e
c
o
m
p
u
t
a
t
i
o
n
a
l
r
e
s
o
u
r
c
e
s
f
o
r
t
a
s
k
s
.
I
n
m
o
r
e
d
e
t
a
i
l
,
t
h
i
s
s
t
e
p
b
a
s
e
d
o
n
r
e
s
o
u
r
c
e
a
v
a
i
l
a
b
i
l
i
t
y
a
n
d
a
p
p
l
i
c
a
t
i
o
n
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
f
o
r
e
a
c
h
t
a
s
k
.
G
e
n
e
r
a
t
e
a
s
a
t
i
s
f
a
c
t
o
r
y
s
o
l
u
t
i
o
n
i
n
s
l
o
t
t
i
m
e
f
o
r
t
a
s
k
s
i
n
c
l
o
u
d
c
o
m
p
u
t
i
n
g
i
s
m
a
i
n
l
y
b
a
s
e
d
o
n
w
o
r
k
f
l
o
w
s
c
h
e
d
u
l
e
r
s
.
T
h
e
w
o
r
k
f
l
o
w
t
e
c
h
n
i
q
u
e
i
n
c
l
o
u
d
c
o
m
p
u
t
i
n
g
i
s
s
h
o
w
n
i
n
F
ig
u
r
e
3
.
Figure
3.
Wo
r
kfl
ow sc
hem
e in cloud
c
om
pu
ti
ng
[
9]
T
h
e
i
n
p
u
t
a
n
d
o
u
t
p
u
t
d
a
t
a
f
i
l
e
s
a
r
e
a
s
s
o
c
i
a
t
e
d
w
i
t
h
e
a
c
h
c
o
m
p
u
t
e
r
p
r
o
g
r
a
m
r
e
p
r
e
s
e
n
t
e
d
b
y
n
o
d
e
s
.
T
h
e
r
o
l
e
o
f
w
o
r
k
f
l
o
w
s
c
h
e
d
u
l
e
r
s
e
a
s
i
l
y
n
o
t
i
c
e
s
w
i
t
h
c
l
o
u
d
c
o
m
p
u
t
i
n
g
i
n
m
o
d
e
r
n
a
p
p
l
i
c
a
t
i
o
n
s
.
I
n
m
o
r
e
d
e
t
a
i
l
,
F
i
g
u
r
e
4
d
e
m
o
n
s
t
r
a
t
e
s
i
n
t
e
r
a
c
t
i
o
n
p
r
o
c
e
s
s
o
f
s
c
h
e
d
u
l
e
r
w
i
t
h
w
o
r
k
f
l
o
w
a
n
d
c
l
o
u
d
r
e
s
o
u
r
c
e
s.
T
h
e
i
n
t
e
r
a
c
t
i
o
n
p
r
o
c
e
s
s
o
f
s
c
h
e
d
u
l
i
n
g
e
n
g
i
n
e
i
s
r
e
p
r
e
s
e
n
t
e
d
i
n
t
h
r
e
e
l
a
y
e
r
s
a
s
s
h
o
w
n
i
n
F
i
g
u
r
e
4
w
h
i
c
h
a
r
e
w
o
r
k
f
l
o
w
,
s
c
h
e
d
u
l
e
r
,
a
n
d
c
l
o
u
d
r
e
s
o
u
r
c
e
s
.
T
h
r
e
e
s
t
e
p
s
a
r
e
a
c
c
o
m
p
l
i
s
h
e
d
i
n
l
a
y
e
r
t
w
o
w
h
i
c
h
a
r
e
t
a
s
k
a
n
a
l
y
s
i
s
,
o
b
j
e
c
t
i
v
e
s
o
p
t
i
m
i
z
a
t
i
o
n
,
a
n
d
t
a
s
k
d
i
s
t
r
i
b
u
t
i
o
n
.
A
l
l
o
c
a
t
e
d
o
f
c
l
o
u
d
r
e
s
o
u
r
c
e
i
n
l
a
y
e
r
t
h
r
e
e
h
e
a
v
i
l
y
d
e
p
e
n
d
o
n
l
a
y
e
r
t
w
o
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
y
br
id
ap
pr
oach
for
sch
e
du
li
ng
applicati
on
.
..
(
A
hm
e
d S.
Ab
da
lk
afo
r)
1391
Fig
ure
4. Sc
he
du
le
r
inter
act
io
n wit
h
cl
ou
d re
so
urces
and
w
orkf
l
o
w
[
9
]
4.
THE
PROPO
SED
ALGO
R
ITHM
A
no
vel
sche
duli
ng
al
gorith
m
is
pr
op
os
ed
in
this
pa
pe
r
to
fi
x
com
m
on
pro
blem
s
of
the
tradit
iona
l
rou
nd
-
r
obin al
gorithm
e
m
plo
ye
d
in cl
oud
c
om
pu
ti
ng
. In
m
or
e
detai
l, it
u
s
es f
irst c
om
e first serve al
gori
thm
to
sche
du
le
al
l
a
rr
ive
d
proces
s
es
in
qu
e
ue
wait
ing
with
f
ixed
ti
m
e
qu
a
ntu
m
.
Howe
ve
r,
wait
ing
ti
m
e
and
tur
naroun
d
ti
m
e
are
crit
eria
ut
il
iz
ed
in
c
or
e
work
of
the
pr
opos
e
d
al
gorithm
.
This
al
go
r
it
h
m
is
en
han
c
ed
i
n
this
pap
e
r
by
re
-
ar
rangin
g
processes
via
sla
pp
e
d
pri
or
it
y
file
d.
Wh
e
n
pr
oce
sses
arri
ve
at
read
y
q
ue
ue,
the
highest
pr
i
or
it
y
is
al
located
to
the
lowe
s
t
pr
oces
s
value
of
burst
ti
m
e.
In
this
case,
al
l
arr
ive
d
proces
ses
at
read
y
que
ue
will
yi
el
d
to
re
-
a
rr
a
ng
i
ng
t
echn
i
qu
e
.
H
oweve
r,
the
ne
w
al
gorithm
play
s
i
m
po
rtant
ro
le
in
enh
a
ncin
g
ove
rall
pe
rfor
m
ance
an
d
t
hro
ughput
rate
of
c
loud
c
om
pu
ti
ng.
Th
us,
it
re
du
ce
s
the
am
ou
nt
of
wait
ing
ti
m
e and tu
rn
a
rou
nd
tim
e. Th
e fl
owchar
t
of the
pro
po
s
e
d al
gorith
m
is show
n
in
Figure
5.
Figure
5. Bl
oc
k diag
ram
o
f
pro
posed
alg
or
it
hm
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
2
,
A
pr
i
l 202
0
:
1387
-
1397
1392
The pri
nciple
work of t
he pr
opos
e
d
sc
he
du
l
ing
tec
hn
i
que i
s sho
wn in
t
he
al
gorithm
.
Algorithm
1
th
e Pro
po
se
d Sc
hem
e
Inp
ut:
N,
BT
, AT,
Pr
i
or
it
y, T
Q
Ou
t
pu
t:
Gan
tt
C
har
t,
A
WT, AT
AT
1:
Arrange all
th
e
pro
ces
ses in
R
Q
in
or
der
of t
heir hig
hest
pr
i
or
it
y t
o
t
he
lo
w
est
v
al
ue
of BT
.
2:
Wh
il
e R
Q
is
not em
pty
3:
Fo
r
I =1
to N
4:
Ca
lc
ulate
the
BT o
f
the
pro
c
esses;
5:
Delet
e the
proc
ess from
RQ
if
the BT=
0;
6:
If
t
he new
pr
oc
ess ar
rive
d
R
Q
go to Ste
p 2;
7:
Dr
a
w Gantt C
ha
rt and Cal
c
ulate
AWT,
AT
A
T;
8:
En
d wh
il
e.
Th
e
ste
ps
of
the
pro
pose
d
al
gorithm
are
li
ste
d
in
Al
gorithm
1.
T
he
a
lgorit
hm
sta
rts
with
the
ra
ndo
m
init
ia
li
zation
of tasks:
St
ep
1:
Alloca
te
CPU
t
o
e
very
process
in
th
e
r
ound
-
robin
appr
oach,
new
pr
i
or
it
ie
s
are
a
ssign
e
d
t
o
al
l
a
rr
ive
d
processe
d
acc
ordi
ng
t
o
g
ive
t
he
highest
pri
ori
ty
to
the
lo
we
st
value
of
bu
r
st
tim
e
with
the
sam
e
t
i
m
e
qu
antum
tim
e.
Then
eac
h
process
gets
the
co
ntr
ol
of
t
he
CP
U
acc
ordin
g
to
the
ne
w
pr
io
riti
es
ba
sed
on
the
rem
ai
ning
CPU
bursts i
n ca
se the
rea
dy
qu
e
ue
is
not re
ached t
o
t
he n
ul
l
St
ep
2:
Af
te
r
f
inishin
g
the
fir
st
ste
p,
cal
cula
te
the
burst
ti
m
e
of
al
l
pr
oc
esses
an
d
re
-
ar
range
d
the
rem
ai
nin
g
burst ti
m
e w
it
h
the
new pri
ori
ti
es unti
l t
he
val
ue of
a
burst ti
m
e reach
e
d
ze
r
o.
5.
SIMULATI
O
N RESULT
A
ND D
I
SCUS
S
ION
To
e
valuate
the
pe
rfo
rm
ance
of
t
he
pro
po
s
ed
syst
em
,
var
i
ou
s
job
processes
a
re
app
li
ed
to
the
tradit
io
nal
rou
nd
-
r
obin
al
gorithm
and
t
he
propose
d
al
gorithm
.
All
these
sce
nar
i
os
are
em
plo
ye
d
unde
r
env
i
ronm
ent
of
virt
ual
bo
x
of
net
f
ram
ewo
r
k.
In
m
or
e
deta
il
,
five
of
job
s
processes
w
hic
h
are
P1,
P
2,
P
3,
P
4,
p5
are
integ
rat
ed
with
CP
U
burst
ti
m
e.
Thes
e
value
s
of
bur
st
tim
e
are
25,
30,
87,
13,
20
resp
ect
ively
.
I
n
this
pap
e
r,
va
rio
us
pr
i
or
it
y
values
are
util
iz
ed
whic
h
are
2,
3,
4,
0,
1
r
especti
ve
ly
.
The
fixed
ti
m
e
qu
antum
in
this
pro
po
sal
is
20
m
s w
it
h
equ
al
and d
i
ff
e
ren
t a
rr
ival t
im
e in ready q
ue
ue
as
sh
ow
n
in
Ta
ble
1.
Table
1.
Proce
sses
s
pecifica
ti
on
s
Proces
s Na
m
e
Arr
iv
al
Ti
m
e
Bu
rst T
i
m
e
Priority
P
1
0
25
2
P
2
0
30
3
P
3
0
87
4
P
4
0
13
0
P
5
0
20
1
The
pe
rfor
m
ance
m
et
rics
are
consi
der
e
d
a
ve
ry
i
m
po
rtant
f
act
or
in
evalua
t
ing
the
pro
po
s
ed
syst
e
m
.
These
m
et
rics,
su
ch
as
t
urna
rou
nd
ti
m
e
an
d
wait
in
g
tim
e
[14].
A
ver
a
ge
of
tu
rn
a
r
ound
tim
e
(TA
T)
c
an
be
cal
culat
ed
as:
Av
e
ra
ge ATA
= Ʃ (
Ti
–
AT
of P
i)/
N
(1)
A
ve
ra
ge of wa
it
ing
ti
m
e (W
T
)
ca
n be calc
ul
at
ed
as:
Av
e
ra
ge WT=
Ʃ (T
AT of
Pi
–
BT o
f Pi
)/
N
(2)
In this
pap
e
r, t
wo sce
nar
io
s a
re a
pp
li
ed
to
te
st
the
ef
fici
enc
y of t
he pr
opos
ed
syst
em
.
O
n
on
e
ha
nd,
we
assum
e
al
l
pr
oble
m
a
rr
ival
at
the
sam
e
tim
e
in
scenari
o
I.
I
n
this
case,
the
tradit
io
nal
rou
nd
-
r
obin
i
s
buil
t
on
distrib
ution
of
the
tim
e
qu
antu
m
a
m
on
g
the
current
proces
ses
i.e.
the
al
gorithm
handles
al
l
pro
cesses
with
out
featur
e
of
the
pr
i
or
it
y.
The
processes
distr
ibu
ti
on
of
tra
di
ti
on
a
l
rou
nd
-
r
obin al
gorithm
is
sh
own
in
F
ig
ur
e
6
.
A
ver
a
ge of
WT: 74.
4m
s; Av
erag
e
of T
AT:
109.4m
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
y
br
id
ap
pr
oach
for
sch
e
du
li
ng
applicati
on
.
..
(
A
hm
e
d S.
Ab
da
lk
afo
r)
1393
Figure
6.
Ga
ntt
c
ha
rt of
t
ra
diti
on
al
R
ound
-
r
obin
a
l
gorithm
On
the
ot
her
ha
nd,
to
e
valuat
e
the
perf
or
m
a
nce
of
the
pro
po
s
ed
al
gorith
m
,
we
ap
plied
to
the
sam
e
scenari
o
of
t
he
pr
e
vious
jobs.
In
ad
diti
on,
th
e
highest
pr
i
or
i
ty
value
at
ta
ch
ed
to
t
he
lo
wes
t
value
of
bur
st
tim
e
resp
ect
ively
.
A
dd
i
ng
proc
ess
heav
il
y
de
pe
nds
on
a
rr
ival
ti
m
e
of
j
obs
at
r
eady
que
ue
tha
t
hav
e
the
sam
e
TQ
.
The
G
antt
cha
rt
of
t
he
pro
po
sed
a
lg
or
it
hm
is
sh
ow
n
Fig
ure
7.
A
ver
a
ge
of
WT:
58.
4ms;
Av
e
rag
e
of
TAT:
93.4
m
s.
Figure
7. Ga
ntt cha
rt of
pro
posed
al
gorithm
Howe
ver,
Figure
8
show
s
the
per
f
or
m
ance
com
par
ison
of
j
obs
at
read
y
qu
e
ue
s
that
ha
ve
the
sam
e
arr
ival
ti
m
e
.
T
he
unit
es
of
sa
ved
ti
m
e
fo
r
th
e
aver
a
ge
of
WT
an
d
T
AT
after
ap
plyi
ng
the
pro
posed
a
lgorit
hm
that
is
s
how
n
i
n
T
a
ble
2
.
Whereas,
the
proc
ess
will
hav
e
r
epeate
d
in
sce
nar
i
o
II
agai
n
t
o
c
onfirm
eff
ic
ie
ntly
of
the
pr
opos
e
d
sche
du
li
ng
al
gorithm
.
In
this
case,
we
assu
m
e
t
hat
j
ob
s
at
a
read
y
qu
e
ue
with
var
io
us
a
rr
ival
tim
e
s
. Tab
le
3
sh
ows
pro
c
ess
sp
eci
ficat
io
n
s
f
or f
ive
jo
bs
in
t
he
qu
e
ue.
Figure
8
.
Per
f
orm
ance co
m
par
iso
n for sce
na
rio
Table
2.
Re
s
ults f
or
S
ce
nar
io
I
Perf
o
r
m
an
ce
Attribu
te
Tr
ad
itio
n
al
Ro
u
n
d
-
rob
in
Sch
ed
u
lin
g
Prop
o
sed
Algo
rith
m
Re
m
ark
Av
erage
W
aiti
n
g
Ti
m
e
7
4
.4
ms
5
8
.4
ms
1
6
Unites
of
Sav
ed
T
i
m
e
Av
erage
Turn
arou
n
d
Ti
m
e
1
0
9
.4
ms
9
3
.3
ms
1
6
.1 Un
ites
Sav
ed
T
i
m
e
Table
3.
Proce
sses S
pecifica
ti
on
s
Proces
s Na
m
e
Arr
iv
al
Ti
m
e
Bu
rst T
i
m
e
Priority
P
1
1
25
2
P
2
3
30
3
P
3
5
87
4
P
4
2
13
0
P
5
0
20
1
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
2
,
A
pr
i
l 202
0
:
1387
-
1397
1394
The
num
ber
of
job
s
with
their
slot
ti
m
e
fo
r
tr
aditi
on
al
rou
nd
-
r
ob
i
n
is
s
how
n
in
F
ig
ure
9
.
Av
e
ra
ge
of
WT:
53.
6m
s;
Av
e
ra
ge
of
T
A
T:
88.6
m
s.
The
nu
m
ber
of
j
ob
s
with
their
slo
t
tim
e
fo
r
th
e
pro
posed
al
gorithm
is
sh
ow
n
in
F
ig
ure
10.
Wh
erea
s
,
Av
e
ra
ge
of
WT:
49.
6m
s;
Av
e
ra
ge
of
T
AT:
84.
6m
s.
P
erfor
m
ance
co
m
par
ison
in
va
rio
us
a
rr
i
val
tim
e
s
is
sh
own
i
n
Fig
ure
11
.
T
o
discusse
s
the
propose
d
al
gorithm
,
w
e
nee
d
to
c
ompare
ou
r
al
gorithm
with
tradit
ion
al
to
cal
culat
e
the
eff
ic
ie
nt
rate
of
the
al
gorithm
.
Acc
ordin
g
to
the
res
ult,
we
can
easi
ly
no
ti
ce
t
hat
the
propos
ed
al
gorithm
m
or
e
eff
ic
ie
nt
fr
om
the
tradit
ion
al
al
go
rith
m
.
Ou
r
pro
po
sal
is
app
li
ed
w
it
h
th
e prev
i
ou
s
alg
ori
thm
[
12
,
14
-
17]
to
m
easur
e i
ts effici
ency.
Figure
9.
Ga
ntt
c
ha
rt for
roun
d
-
r
obin
a
lg
or
it
hm
Figure
10.
Ga
nt
t chart
of
t
he p
rop
os
ed
alg
or
it
hm
Figure
11.
Per
f
or
m
ance co
m
par
iso
n for sce
na
rio II
a.
Im
pr
ov
e
d r
ound
-
robin
sc
hedu
li
ng
in
cl
ou
d
c
om
pu
ti
ng
[
12
]
So
m
e
per
form
ance
m
et
rics
a
re
util
ise
d
to
evaluate
ef
ficacy
of
ne
w
sche
du
li
ng
m
et
ho
d
fo
r
cl
oudi
ng
com
pu
ti
ng
that
pro
posed
[
12]
.
T
hu
s
,
Fig
ur
e
12
s
hows
im
prov
e
rate
of
rou
nd
-
r
obin.
A
ve
r
age
of
WT:
42
.4
m
s
;
Av
e
ra
ge of TA
T: 6
2.4m
s.
Figure
12.
Ga
nt
t chart
of
im
prov
e
d r
ound
-
robin sche
duli
ng
b.
En
han
ce
d rou
nd
-
robin
[1
4]
In
Fi
gure
13,
aver
a
ge
of
WT
and
a
ver
a
ge
of
TA
T
are
cal
culat
ed
to
m
easur
e
im
pr
ov
i
ng
rate
of
the ro
und
-
r
ob
i
n [14].
Av
e
ra
ge
of
WT: 64.
2m
s
; Av
era
ge o
f
T
AT:
112.4m
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
y
br
id
ap
pr
oach
for
sch
e
du
li
ng
applicati
on
.
..
(
A
hm
e
d S.
Ab
da
lk
afo
r)
1395
Figu
re
13. G
ant
t chart
of
en
ha
nced r
ound
-
r
obin
c.
In
te
ger
p
rogr
a
m
m
ing
[
15]
In
a
dd
it
io
n,
a
ver
a
ge
of
W
T
and
ave
ra
ge
of
T
AT
are
r
equ
i
red
to
te
st
integer
pro
gr
a
m
m
ing
f
or
the sc
hedulin
g m
et
ho
d
t
hat s
how
n
i
n
Fi
gure
14 [1
5].
A
ve
ra
ge of
W
T:
36m
s; Avera
ge of
TAT:
84.6
m
s.
Figure
14.
Ga
nt
t chart
of
i
nteg
er
program
m
ing
robin
alg
or
it
hm
d.
Im
pr
ov
e
d r
ound
-
robin
alg
or
it
hm
[
17
]
Nayak
et
al
,
ha
ve
propose
d
sche
du
li
ng
al
gorithm
to
i
m
pr
ov
e
pe
rfor
m
ance
of
r
ound
-
robin
m
et
hod.
Howe
ver, f
i
gur
e 15 s
hows
ef
fici
ency o
f
this
a
lgorit
hm
.
Av
er
age
of
W
T:
84
m
s
; Av
era
ge o
f
T
AT:
141.6m
s.
Fi
g
ure
15.
Ga
nt
t chart
of
im
prov
e
d r
ound
-
robin
al
gorithm
e.
A
pr
i
or
it
y
-
base
d
rou
nd
-
r
obin
[
16
]
Ra
j
put
et
al
[
16]
,
ha
ve
prese
nted
s
om
e
i
m
p
rovem
ent
proc
ess
on
rou
nd
-
r
ob
i
n
al
gorit
hm
as
sho
wn
in
Figure
16.
A
ve
rag
e
of
WT:
24.4
m
s; Av
era
ge
of TAT
: 3
5m
s.
Figu
r
e
16.
Ga
nt
t chart
of
pri
or
it
y
-
based
r
ound
-
robin
alg
or
it
hm
To
disti
ng
uish
our
w
ork
f
rom
oth
er,
we
ne
ed
com
par
iso
n
stu
dy
with
the
rece
nt
pap
e
rs
at
the
sa
m
e
resear
c
h
area.
Ta
ble
4
sh
ows
c
om
par
iso
n
stu
dy
betwee
n
va
rio
us
pr
opos
e
d
a
lgorit
hm
on
rou
nd
-
r
obin
m
et
ho
d.
The
s
umm
ary
of
th
e
com
par
ison
betw
een
t
he
pre
vious
sc
hed
ulin
g
al
gor
it
h
m
s
with
our
proposal
sc
he
du
l
ing
al
gorithm
is sh
own
i
n
Fi
gure
17.
Table
4
c
om
par
iso
n betwee
n vari
ou
s
im
pr
ove on
rou
nd
-
r
ob
in.
Perf
o
r
m
a
n
ce Attr
ib
u
te
AW
T
AW
T
Af
ter
App
ly
i
n
g
the
p
rop
o
sed
algo
rith
m
ATAT
ATAT
Af
ter
Ap
p
ly
in
g
th
e pro
p
o
sed
alg
o
r
ith
m
San
g
wan
et al
.
[
1
2
]
4
4
.8
ms
4
2
.4
ms
6
4
.8
ms
6
2
.4
ms
Mittal
et
al
.[
1
4
]
6
8
.8
ms
6
4
.2
ms
117
ms
1
1
2
.4
ms
Ku
m
ar
et al
.[
1
5
]
7
1
.6
ms
36
ms
1
1
9
.8
ms
8
4
.6
ms
Rajp
u
t
et al
.[
1
6
]
28
ms
2
4
.4
ms
40
ms
35
ms
Nay
ak
e
t al
.[
1
7
]
8
5
.6
ms
84
ms
1
4
3
.2
ms
1
4
1
.6
ms
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
2
,
A
pr
i
l 202
0
:
1387
-
1397
1396
Fig
ure
17.
C
om
par
ison
betw
een
pr
e
vious al
gorithm
s w
it
h new al
gorithm
6.
CONCL
US
I
O
N
A
novel
sc
heduling
al
gorith
m
fo
r
cl
oud
co
m
pu
ti
ng
is
pr
opose
d
i
n
this
pa
per
t
o
ov
e
rc
om
e
com
m
on
pro
blem
s
of
round
-
r
ob
i
n
al
go
rithm
.
It
has
the
abili
ty
to
cr
eat
e
balance
lo
ad
bet
wee
n
jo
b
pr
ocesses
i
n
cl
oud
com
pu
ti
ng
at
read
y
queue
.
The
pro
pr
ie
ty
value
at
ta
che
d
to
bur
st
tim
e
play
s
a
vital
ro
le
i
n
e
nh
a
nci
ng
perform
ance
of
cl
oud
com
pu
ti
ng
sc
he
du
li
ng.
Wh
en
c
om
par
e
ou
r
pr
opo
s
al
with
the
pr
e
vious
al
gorithm
,
we
can
easi
ly
no
ti
ce
that
the
pro
po
se
d
al
gor
it
h
m
ha
s
a
signi
ficant
eff
ect
on
pe
rfor
m
ance
of
cl
ou
d
com
pu
ti
ng
.
Our
ex
per
im
ental
resu
lt
of
th
e
propose
d
job
s
scheduli
ng
al
gorithm
sh
ow
s
that
the
pr
op
ose
d
schem
es
po
sses
s
ou
tst
a
nd
i
ng
e
nhanci
ng
rates
with
a
re
duct
io
n
in
wait
ing
ti
m
e
fo
r
j
obs
i
n
qu
e
ue
li
st.
I
n
f
uture
w
ork,
t
his
idea
will
app
ly
on
an
ot
her
sc
he
duli
ng
al
gorithm
for
cl
oud
c
ompu
ti
ng
that
ta
r
geted
to
reduc
e
rate
of
wait
ing
ti
m
e
and tu
rn
a
rou
nd tim
e at ready
qu
e
ue wit
h.
RE
FERE
NCE
S
[1]
N.
Kulkar
ni
,
S.
V.
N.
L.
Lalit
h
a
,
and
S.
A
.
Deo
kar
,
“
Real
Ti
m
e
Control
and
M
onit
oring
of
Gri
d
Pow
er
S
y
stem
s
using
Cloud
Com
puti
ng,
”
Inter
nati
onal
Journal
of
El
e
ct
rica
l
and
Computer
Engi
nee
ring
,
vol.
9
,
No.
2,
pp.
941
-
949
,
20
19.
[2]
K.
Bhagc
hand
an
i
and
D.
P.
Augus
ti
ne
,
“
IoT
Bas
ed
Hea
rt
Monito
ring
and
Aler
ti
n
g
Sy
st
em
with
Cloud
Com
puti
ng
and
Mana
ging
The
Tra
ff
ic
for
an
Am
bula
nce
in
India
,
”
Inter
nati
onal
Journal
of
El
e
ct
rica
l
and
Computer
Engi
ne
ering
,
vol
.
9
,
no
.
6
,
pp
.
50
68
-
5074,
2019
.
[3]
H.
G.
Ta
n
i,
and
C.
El
Am
ran
i
,
“
Cloud
C
om
puti
ng
CP
U
A
ll
oca
t
ion
and
Schedu
l
ing
Algorit
hm
s
using
CloudSi
m
Sim
ula
tor,
”
Int
e
rnational
Journal
of
Elec
tric
al
an
d
Computer
Eng
ine
ering
,
vol
.
6
,
no.
4
,
pp
.
1866
-
1879,
2016
.
[4]
W
.
Hus
sain,
F.
K.
Hus
sain,
O.
K.
Hus
sain,
E.
Dam
ia
ni,
and
E.
Chang,
“
Form
ula
ti
ng
and
Mana
g
ing
Viable
SLAs
in
Cloud
Com
puti
ng
from
a
S
m
al
l
t
o
Medium
Servic
e
Provider
'
s
Vi
ewpoint
:
A
State
-
o
f
-
t
he
-
Art
Rev
i
ew,
”
Informatio
n
Syste
ms
,
vol
.
71
,
pp.
240
-
259,
20
17.
[5]
S.
Ghanba
ri
an
d
M.
Othm
an
,
“
A
Priority
B
ase
d
Job
Scheduling
Al
gorit
hm
i
n
Cloud
Com
puti
ng,
”
Proc
edia
Engi
ne
ering
,
vol
.
50
,
pp
.
778
-
78
5,
2012
.
[6]
R.
Nall
akumar
,
“
A
Surve
y
on
Dea
dli
n
e
Constrai
ned
W
orkflow
Schedul
ing
Algorit
hm
s
in
Cloud
Envi
ronm
ent,
”
Inte
rnational
Jo
urnal
of
Computer
Sci
en
ce
Tr
en
ds
and
Technol
ogy
(
IJCST
)
,
vol.
2,
no.
5,
pp
.
44
-
50,
2014.
arXiv
pre
print a
rXiv
:1
409.
7916.
[7]
T.
Dil
lon,
C
.
W
u,
and
E.
Ch
ang,
“
Cloud
Co
m
puti
ng:
Iss
ues
and
C
hal
l
enges
,
”
2010
24th
I
EE
E
In
te
rnation
al
Confe
renc
e
on
A
dvanc
ed
Information
N
et
working
and
Applications,
Aus
tra
l
ia,
pp
.
2
7
-
33,
2
010
.
[8]
W
.
C.
W
u
and
S
.
C.
Yang
,
“
Enh
anc
ing
Im
ag
e
Se
cur
ity
and
P
rivac
y
in
Cloud
S
y
st
em
using
Stega
n
ogra
ph
y
,
”
IEEE
Inte
rnational
Co
nfe
renc
e
on
Con
sum
er
El
ectroni
cs
-
Taiwan (
ICC
E
-
TW)
,
pp.
321
-
322,
June
2017.
[9]
I.
C.
Lopez,
“
Scie
nti
fi
c
W
orkflow
Scheduling
f
or
Cloud
Com
puti
ng
E
nvironments,
”
Doctoral
Thesi
s,
The
Unive
rs
ity
o
f
Sydn
ey
,
2016
.
[10]
Y.
H.
Le
e
,
S.
Le
u,
and
R.
S.
Chang,
“
Im
proving
Job
Schedu
l
ing
Algorit
hm
s
in
a
Grid
Env
ir
onm
ent
,
”
Fut
ur
e
gene
ration
comp
ute
r syste
ms
,
vol
.
27
,
no
.
8
,
pp
.
9
91
-
998,
2011
.
[11]
S.
K.
Panda
and
S.
K.
Bhoi,
“
An
Eff
ec
ti
ve
Round
-
Robin
Algorit
hm
using
Min
-
Max
Dispersion
Mea
sure,
”
Inte
rnational
Jo
urnal
on
Compu
te
r Sc
ie
nc
e
and
Engi
ne
ering
,
20
14.
arXi
v
prepri
nt
arXi
v:1
404.
5
869
.
[12]
P.
Sangwan,
M.
Sharm
a,
and
A.
Kum
ar,
“
I
m
pro
ved
Round
-
robi
n
Schedul
ing
in
Cloud
Com
puti
ng,
”
Adv
an
ce
s
in
Computati
onal
S
ci
en
ce
s and
Tec
hnology
,
vol
.
10
,
no.
4,
pp.
639
-
6
44,
2017
.
[13]
B.
Mani
y
ar
,
an
d
B.
Kana
ni
,
“
Revi
ew
on
Rou
nd
-
Robin
Algorit
hm
for
Ta
sk
Schedul
ing
in
Cloud
Com
puti
ng,
”
Journal
of
Eme
r
ging
Techno
logies and
Inno
vat
i
v
e
R
ese
arch
,
vol
.
2
,
no
.
3
,
2015
.
[14]
S.
Mitt
al,
S.
Sin
gh,
and
R.
Kaur
,
“
Enha
nc
ed
Round
-
robin
Tech
nique
for
T
ask
Schedul
ing
in
C
loud
Com
puti
ng
Envi
ronm
ent
,
In
te
rnational
Jour
nal
of
Engi
n
ee
ri
ng
Re
search
&
T
ec
hnolog
y
,
pp.
5
25
–
529,
2016
.
Evaluation Warning : The document was created with Spire.PDF for Python.